Quantitative ultrasound radiomics in predicting response to neoadjuvant chemotherapy in patients with locally advanced breast cancer: Results from multi-institutional study.

摘要: Background This study was conducted in order to develop a model for predicting response neoadjuvant chemotherapy (NAC) patients with locally advanced breast cancer (LABC) using pretreatment quantitative ultrasound (QUS) radiomics. Methods multicenter involving four sites across North America, and appropriate approval obtained from the individual ethics committees. Eighty-two LABC were included final analysis. Primary tumors scanned clinical system before NAC started. The contoured, radiofrequency data acquired processed whole tumor regions of interest. QUS spectral parameters derived normalized power spectrum, texture analysis performed based on six features gray level co-occurrence matrix. Patients divided into responder or nonresponder classes their clinical-pathological response. Classification machine learning algorithms, which trained optimize classification accuracy. Cross-validation leave-one-out cross-validation method. Results Based outcomes treatment, there 48 responders 34 nonresponders. A K-nearest neighbors (K-NN) approach resulted best classifier performance, sensitivity 91%, specificity 83%, an accuracy 87%. Conclusion QUS-based radiomics can predict acceptable